mmWave Sensing for Detecting Movement Through Thermoplastic Masks During Radiation Therapy Treatment

Precision in radiation therapy relies on immobilization systems that limit patient motion. Thermoplastic masks are commonly used for this purpose, but subtle voluntary and involuntary movements such as jaw shifts, deep breathing, or eye squinting may still compromise treatment accuracy. Existing motion tracking methods are limited: optical systems require a clear line of sight and only detect surface motion, while X-ray-based tracking introduces additional ionizing radiation. This study explores the use of low-power, non-ionizing millimeter-wave (mmWave) sensing for through-mask motion detection. We characterize the RF properties of thermoplastic mask material in the 28-38 GHz range and perform motion detection using a 1 GHz bandwidth centered at 28 GHz. We use a frequency-domain system with horn antennas in a custom-built anechoic chamber to capture changes in the amplitude and phase of transmitted RF waves in response to subtle head and facial movements. These findings lay groundwork for future real-time through-mask motion tracking and future integration with multi-antenna systems and machine learning for error correction during radiotherapy.

💡 Research Summary

The paper investigates the feasibility of using low‑power, non‑ionizing millimeter‑wave (mmWave) sensing to detect subtle head and facial movements through the thermoplastic immobilization masks commonly employed in radiation therapy. The authors first characterize the electromagnetic transparency of a clinical‑grade polycaprolactone mask in the 28–38 GHz band using a vector network analyzer (VNA) and horn antennas. Measurements reveal a maximum insertion loss of 5.5 dB at 28 GHz with a 50° incidence angle, confirming that the mask introduces only modest attenuation and phase distortion.

A controlled laboratory setup is then built: a bistatic configuration with two 15 dBi horn antennas placed 75 cm apart, providing a 30° beamwidth that fully illuminates a human head. The system operates in a custom tabletop anechoic chamber to suppress multipath reflections. Baseline stability is verified with a static plastic mannequin behind the mask, showing amplitude fluctuations below 0.1 dB and phase variations under 5°, establishing a reliable reference for subsequent motion experiments.

Human subject tests involve a healthy adult volunteer performing a series of choreographed motions—eye squinting, smiling, yawning, and small left/right head rotations—while seated behind the mask. The VNA sweeps a 1 GHz bandwidth centered at 28 GHz (101 frequency points, 1.7 s sweep time, 50 Hz IF bandwidth), capturing complex S21 parameters for each motion. Compared with the static baseline, eye squint and smile produce amplitude changes of 0.2–0.4 dB and phase shifts of 8–12°, yawning yields 0.5–0.7 dB and 14–18°, and head rotations generate the largest responses (0.6–0.9 dB, 15–22°). All observed variations exceed the baseline noise by a comfortable margin, demonstrating that a single bistatic link can reliably detect sub‑centimeter motions through the mask.

Link‑budget analysis indicates an overall transmission loss of roughly 45 dB, which remains within the dynamic range of typical low‑power mmWave transmitters (≤10 mW). The authors argue that this loss, combined with the measured signal‑to‑noise ratio, is sufficient for real‑time operation once the system is transitioned from frequency‑swept measurements to continuous‑wave or FMCW radar architectures, which can provide millisecond‑scale update rates.

Finally, the paper outlines a roadmap toward a clinical implementation. The envisioned system integrates a circumferential mmWave antenna array around the treatment gantry, enabling multi‑view beamforming and spatial resolution sufficient to differentiate global patient shifts from localized jaw or facial motions. Machine‑learning classifiers would process the amplitude and phase data to identify motion types and feed motion estimates back to the treatment control system for gating, beam adaptation, or couch corrections. The current work serves as a proof‑of‑concept, establishing that thermoplastic masks do not preclude mmWave sensing and that even a simple single‑link setup can detect clinically relevant motions, paving the way for non‑invasive, radiation‑free intrafraction monitoring in radiotherapy.